Evaluated 71 investigations of MMPI usage in prison work. Additional studies were examined to provide a methodological basis for the comparisons of the research, which were made within sections on (a) sampling procedures, (b) sources of variance and their effects on test results, (c) protocol validity, and (d) methods of profile interpretation. Several methodological shortcomings and various differences in procedure across studies limit the generalizability of the findings. Research in the hostile-assaultive section, however, has produced preliminary MMPI indicators for a type of violently aggressive behavior pattern that is otherwise difficult to detect. Other areas in which the MMPI shows promise include homosexuality, recidivism, and the classification of psychopathologic behavior. More research is needed in the areas of institutional adjustment and suicide. Recommendations for future investigations prescribe adequately controlled sampling procedures, modifications in the interpretation of protocol validity, investigation of certain methodological questions in their own right, consideration of more than one aspect of profile data, the use of base-rate probabilities in predictive studies and the pursuit of longitudinal studies with thorough follow-up procedures. (3½ p ref)